Transfer effects in mergers & acquisitions : the role of strategic motives

Transfer Effects in Mergers & Acquisitions:

The Role of Strategic Motives

Master Thesis

Student: Mathijs van Itterzon/ Student No. 11324074 MSc. Business Administration – Strategy track

University of Amsterdam, Faculty of Economics and Business

Supervisor: Dhr. MSc. B. Silveira Barbosa Correia-Lima

University of Amsterdam, Amsterdam Business School

STATEMENT OF ORIGINALTIY

This document is written by, Mathijs van Itterzon who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business responsible solely for the supervision of completion of the work, not for the content.

TABLE OF CONTENTS

ABSTRACT ... 4

INTRODUCTION ... 5

THEORY AND HYPOTHESIS ... 9

Organizational learning ... 9

Acquisition experience and acquisition performance ... 11

Transfer theory of learning ... 13

Literature gap and research question ... 15

Non-linear relationship of acquisition experience and acquisition performance ... 16

Firm age ... 19 Strategic motives ... 21 METHODOLOGY ... 27 Sampling strategy ... 27 Dependent variable ... 28 Independent variable ... 30 Moderating variables ... 31 Control variables ... 33 Statistical model ... 36 RESULTS ... 38

Descriptive statistics and correlation analysis ... 38

Robust regression analysis ... 39

DISCUSSION ... 44

Major findings ... 44

Contributions of this study ... 48

Limitations and implications for future research ... 50

CONCLUSION ... 52

REFERENCES ... 53

ABSTRACT

This study builds on transfer learning theory and examines the moderating effect of strategic motives on the impact between similar acquisition experience and deal performance. Consistent with this theory I argue that it is important to consider the role of acquisition experience heterogeneity based on the strategic rationale in transfer effects. More precise, I hypothesize that the strategic motives behind the deal are not alike and influences the appropriateness of exploiting or disregarding prior acquisition experience to the new acquisition, depending on the type of resources or capabilities sought with the event. Using a unique M&A database with merger motives, this research suggest that tangible (i.e. manufacturing plant) and intangible resources (i.e. patents) allow for appropriate organizational behaviour and consequently do not offset the positive transfer effects of learning between prior acquisition experiences and deal performance such that the positive effect of transfer learning remains the same. In contrast, the type of resource among intangibles does impact this relationship as such that the positive transfer effect of learning becomes weaker under human dependent resources and stronger for human independent resources. Contrary to insights from behavioural learning theory, this study did not find support for an inverted u-shape relationship, suggesting that learning curves in strategic settings are contingent upon the industry under scrutiny. Taken together these results enrich our understanding of organization acquisition experience by providing novel insights about the conditions under which negative transfer occurs.

Key words: strategic motives, transfer effects, acquisitions, M&A, organizational experience, deal performance, behavioural learning theory, transfer theory of learning

INTRODUCTION

For centuries the belief is that learning has a positive effect on the preceding activities. In the following quote: “Give a man a fish and you feed him for a day. Teach a man how to fish and you feed him for a lifetime” the Chinese philosopher and teacher Confucius (551BC-475BC) already mentioned the importance of learning. In research on organization learning, the traditional view of learning curves is long seen as the status quo for firms that predicts positive returns to experience (Barkema & Schijven, 2008; Finkelstein & Haleblian, 2002; Haleblian & Finkelstein, 1999; Hayward, 2002). This perspective entails an important aspect of learning by proposing that the outcome (i.e. unit cost) is a function of a person’s progress in gaining experience from a repetitive task conducted over time (Darr, Argote, & Epple, 1995). There is common ground among researchers that organizations learn by encoding experience from historical conditions into routines that guide their future behaviour (Cyert & March, 1963; Levitt & March, 1988). However, scholars have argued that the insights from traditional experiential learning curves are restricted to the manufacturing context and are not generalizable to strategic settings because the conditions are not alike (Barkema & Schijven, 2008; Finkelstein & Haleblian, 2002; Haleblian & Finkelstein, 1999). The execution of an acquisition is not a matter of simply reapplying previous acquisition experience because deals are not all alike. It can vary among several dimensions such as the type of industry, country, size, hostility, type of financing, technological proximity and strategic motives, which translates into deal experience heterogeneity (Ahuja & Katila, 2001; Barkema & Schijven, 2008; Bower, 2001; K. M. Ellis, Reus, Lamont, & Ranft, 2011; Finkelstein & Haleblian, 2002; Haleblian & Finkelstein, 1999; Hayward, 2002; Hitt, Harrison, Duane Ireland, & Best, 1993; Jo, Park, & Kang, 2016; Nadolska & Barkema, 2007)

Therefore, the effect of organizational experience from previous conditions on a subsequent event is more difficult to predict for acquisitions and as several studies elucidate, does not necessarily ensure enhanced deal performance (K. M. Ellis et al., 2011; Haleblian & Finkelstein, 1999; Hayward, 2002; Lubatkin, 1983).

Substantial empirical evidence from research on the relationship between organizational experience and acquisition performance confirm conflicting results (K. M. Ellis et al., 2011; Finkelstein & Haleblian, 2002; Haleblian & Finkelstein, 1999; Hayward, 2002; Nadolska & Barkema, 2007, 2007, 2014). Studies show positive results (Barkema, Bell, & Pennings, 1996; K. M. Ellis et al., 2011), non-significant (Lubatkin, 1982; Zollo & Singh, 2004), U-shaped (Haleblian & Finkelstein, 1999; Nadolska & Barkema, 2007), inverted U-shaped (Hayward, 2002; Jo et al., 2016) and negative results (Uhlenbruck, Hitt, & Semadeni, 2006). To explain these mixed empirical findings between acquisition experience and deal performance scholars build on the fundamentals of transfer theory, derived from psychology research (K. M. Ellis et al., 2011; Finkelstein & Haleblian, 2002; Haleblian & Finkelstein, 1999; Jo et al., 2016; Nadolska & Barkema, 2007). This theory explains that there is a relationship between prior accumulated experience and the transfer of that experience to the following strategic event. This transfer effect can be either positive or negative (Cormier & Hagman, 1987 in K. M. Ellis et al., 2011). Positive transfer occurs if there is similarity between the transfer source and the transfer target, which allows the firm to recognize common factors and appropriately draw upon prior experience in developed routines, exploiting it to the new acquisition (Salomon & Perkins, 1989; Woodworth & Thorndike, 1901). The outcome from transfer effects of learning is thus contingent upon the appropriateness of exploiting or disregarding prior acquisition experience to the new acquisition by the firm. However, generalizing or

discriminating prior experience from the new acquisition is challenging in strategic settings due to the complexity and experience heterogeneity among deals (Barkema & Schijven, 2008; Haleblian & Finkelstein, 1999; Hayward, 2002)

The current body of knowledge acknowledges the importance of similarity and experience heterogeneity given the numerous studies on this field. Studies tested experience heterogeneity by industry type, country, deal size specificity and technological proximity (Ahuja & Katila, 2001; K. M. Ellis et al., 2011; Finkelstein & Haleblian, 2002; Haleblian & Finkelstein, 1999; Hayward, 2002; Jo et al., 2016; Nadolska & Barkema, 2007, 2014). More important for theorizing in this study is that a growing stream of scholars argues that acquisition experience heterogeneity based on the strategic rational could affect post-merger performance and that understanding the (strategic) motives underlying them is important to successfully implement different types of acquisitions (Ahuja & Katila, 2001; Berkovitch & Narayanan, 1993; Bower, 2001; Hayward, 2002; Jo et al., 2016). Even though existing studies recognize this type of heterogeneity and address the importance, they did not empirically test the effects of strategic motives on the impact between acquisition experience and deal performance.

Using insights from transfer theory I argue that the strategic motive may influence the appropriateness of exploiting or disregarding prior acquisition experience to the new acquisition, depending on the type of resources or capabilities sought with the event. For example, the similarities between tangible resources (i.e. property plant and equipment) and intangible resources (i.e. patents) may be evident. However, appropriate discriminating the underlying acquisition routines for intangible assets such as a brand or intellectual property may be more challenging (Coff, 1999; Gerbaud & York, 2007; Hall, 1992, 1993; Newmeyer, Swaminathan, & Hulland,

2016; Ranft & Lord, 2002). Despite the valuable contributions of earlier research that tested the effects of experience relatedness in several contingencies (i.e. industry, country, size, technological proximity), I thus propose that we can extend our current insights if we also incorporate deal experience heterogeneity based on the strategic motive. In this study I will explore the moderating effect of the strategic rationale for the acquisition on the impact between acquisition experience and deal performance. The contributions in this paper are two-fold. In the first place I theoretically extend the current insights of negative transfer learning effects, by unravelling the moderating role of acquisition experience heterogeneity based on the strategic rationale on the impact between prior acquisition experience and deal performance. This enriches our understanding of the predicted effects of acquisition experience on a subsequent event from the perspective of strategic motives. Conversely, it clarifies how organizations determine the appropriateness of exploiting or disregarding prior acquisition experience to the new acquisition, depending on the strategic resources or capabilities sought with the event. In the second place I elucidate how the curvilinear relationship between acquisition experience and deal performance is contingent upon the industry under scrutiny and could explain the mixed empirical findings in the literature.

THEORY AND HYPOTHESIS

This chapter aims to clarify the constructs used in this study. Section one explains how organizations learn from experience. The second section elaborates the mixed findings from the literature on the relationship between experience and deal performance. Section three explains the fundamentals of transfer learning theory. Section four elaborates on existing studies, while working towards the research gap and research question. Finally, section five, six and seven develops the hypotheses. The definition Merger and Acquisitions (M&A) entails a wide variety of transactions with different traits (i.e. financing, motive, structure, attitude, scope) and distinct implications for both acquirer and target (J. Ellis, Moeller, Schlingemann, & Stulz, 2011; Netter, Stegemoller, & Wintoki, 2011). Although M&A technically differ from each other, this study will not make a demarcation between the two of them. Instead the definitions M&A and acquisitions will be used interchangeable, referring to the same strategic setting (Haleblian, Devers, McNamara, Carpenter, & Davison, 2009).

Organizational learning

From an organizational learning perspective, organizations are seen as learning by encoding experiences from history into routines that guide future behaviour (Cyert & March, 1963; Levitt & March, 1988). Here routines are defined as the forms, rules, procedures, processes, strategies and technologies around which organizations are constructed and through which they operate (cf. Levitt & March, 1988, p. 320). In strategic settings these routines encompasses for example the negotiation, due diligence, valuation and post-merger integration (Ranft & Lord, 2002; Zander & Kogut, 1995; Zollo & Singh, 2004). Insights from behavioural learning theory explain that organizations learn from the consequences of antecedent-behaviour (Haleblian &

Finkelstein, 1999; Woodworth & Thorndike, 1901). Antecedent conditions are referred to as similarity or dissimilarity of focal acquisition experience and behaviour the degree to which firms are able to discriminate between current acquisition and past acquisition experience. If the antecedent condition is similar to past experience, the firm behaviour – action – should be to generalize past experience, which translates into enhanced firm performance (Woodworth & Thorndike, 1901). Feedback from this ‘reward’ consequently acknowledges the appropriateness of the past experience and therefore remains embedded in the organizational routines (Levitt & March, 1988). Likewise, firms that fail to discriminate between dissimilarity of the antecedent condition and past situations will experience ‘punishment’ in terms of poor firm performance. As a result, firms change their behaviour by adjusting the organizational – acquisition – routines according to the new insight. Research point out, two other scenarios for firms those are able to discriminate. Organizational behaviour that inappropriately disregard the antecedent experience from the past experience and vice-versa, appropriate discrimination of dissimilar focal experience from previous experience (Haleblian & Finkelstein, 1999). The predicted consequence for firm performance in these situations is neutral. Thus, the antecedent-behaviour-consequence predicts that firms can expect positive, negative and neutral effects, depending on the organizational ability to discriminate similarities.

However, discrimination is challenging in strategic settings due to the complexity and experience heterogeneity among deals (Barkema & Schijven, 2008; Haleblian & Finkelstein, 1999; Hayward, 2002). Acquisition experience can vary among several dimensions such as, but not limited to, the type of industry, country, size, hostility, type of financing, technological proximity and strategic motives (Ahuja & Katila, 2001; Barkema et al., 1996; Bower, 2001; K. M. Ellis et al., 2011; Finkelstein &

Haleblian, 2002; Haleblian & Finkelstein, 1999; Hayward, 2002; Hitt et al., 1993; Jo et al., 2016; Nadolska & Barkema, 2007). These dimensions are referred to as ‘surface features’, whereas the underlying acquisition routines such as the negotiation or integration are referred to as ‘structural features’ (Haleblian & Finkelstein, 1999). The ability to discriminate is furthermore affected by the level of expertise from the decision maker at the strategic apex (Mintzberg, 1979). Research shows that novice firms are more prone than experts firms to inappropriate generalization (Finkelstein & Haleblian, 2002; Haleblian & Finkelstein, 1999; Hitt et al., 1993; Lubatkin, 1983). Given their infancy and limited experience, they rely more on the surface (‘visible’) similarities between the focal acquisition and past conditions and not on the underlying structural (‘metaphysical’) differences.

Acquisition experience and acquisition performance

In general, acquisitions appear in cycles of seven years. Historical data show that each cycle is associated with a significant increase in deal value creation. According to Thomson Reuters (2016) global deal value hit an all-time ballooning record in 2015 of $4.7 trillion. However, these strategic projects are often unsuccessful resulting in value destruction for the acquirer (Jensen, 1986; Porter, 1987). Still firms execute acquisitions for reasons such as the desire for ‘empire building’, efficient internal markets, market imperfections, operating economies, value creation, bargaining power or hubris (Jensen, 1986; Montgomery, 1994; Porter, 1987; Williamson, 1999). Despite the negative results, profound research shows that more experienced acquirers perform better than their peers that lack acquisition experience (Hitt et al., 1993; Lubatkin, 1983). While learning from prior experience seems intuitive, substantial empirical evidence from research on the relationship between organizational

experience and acquisition performance shows mixed results (Barkema & Schijven, 2008; J. Ellis et al., 2011; Nadolska & Barkema, 2007, 2014).

For example, Finkelstein & Haleblian (2002), Haleblian & Finkelstein (1999) and Hayward (2002) found empirical evidence that industry related experience positively relates to deal performance. These insights predict that commonalities between antecedent (focal) acquisition and past acquisition experience in the same four-digit Standard Industry Classification (SIC) industry enhance deal performance in terms of synergies. Likewise, existing literature is full of evidence that unrelated (conglomerate) acquisitions lead to value destruction due to absence of strategic fit, as opposed to related deals (e.g. Montgomery, 1994). Other scholars found evidence that cross-border acquisitions negatively impacts deal value creation if cultural differences are not acknowledged (Barkema et al., 1996; Hitt, Hoskisson, & Kim, 1997; Nadolska & Barkema, 2007). Firms are able to reap the benefits from acquisitions if they exploit previous experience in related countries and in countries that share the same cultural distance. These empirical results show that learning effects are relative stronger for related cultural and allows for appropriate generalizing and transferring of existing acquisition routines. This translates into i.e. faster post-merger integration. K. M. Ellis et al. (2001) found support that small-related acquisition experience cannot be simply transferred to large focal acquisitions due to the associated complexity (K. M. Ellis et al., 2011; Hayward, 2002; Woodworth & Thorndike, 1901). In fact, dissimilarity does not allow acquisition routines to be transferred one on one because of the post-merger integration complexity associated to large acquisitions. Consequently large deals thus hinder deal performance in terms of a longer period of integration, whereas small related acquisition do not.

Results from studies in high-technological industries indicate that the dissimilarity of the antecedent technological knowledge base compared to previous acquisitions cannot simply be applied due to asset intangibility (Ahuja & Katila, 2001; Bena & Li, 2014; Datta & Roumani, 2015; Hitt et al., 1997; Jo et al., 2016; Kogan, Papanikolaou, Seru, & Stoffman, 2016; Sears & Hoetker, 2014). Acquisition routines are dissimilar due to the difficulty of valuation for intellectual property and integration of human dependent knowledge competencies (Gerbaud & York, 2007; Mizik & Jacobson, 2003; Ranft & Lord, 2002; Zander & Kogut, 1995). Thus, experience only allows for generalization for related deals that share the same technological proximity, translating into deal value creation from synergies. These empirical results thus predict enhanced deal performance for related (similar) deals but also implies that firms accumulate similar experience-based acquisition knowledge in their organizational routines, which allows for faster learning from that experience (Haleblian & Finkelstein, 1999; Nadolska & Barkema, 2007). The central insight from these findings is that experience does not necessarily improve deal performance.

Transfer theory of learning

To explain the mixed empirical findings between acquisition experience and deal performance scholars build on the fundamentals of transfer theory, derived from psychology research (K. M. Ellis et al., 2011; Finkelstein & Haleblian, 2002; Haleblian & Finkelstein, 1999; Hayward, 2002; Jo et al., 2016; Nadolska & Barkema, 2007). Woodworth & Thorndike (1901) elucidate in their seminal work that transfer of learning is the degree to which antecedent and prior conditions are similar. In the context of transfer theory, learning can be viewed as the process that consist of a ‘flow’ of knowledge from one setting being applied in another and experience as the

‘stock accumulation’ of that learning (Dierickx & Cool, 1989). The notion is that there is a relationship between prior accumulated experience and the transfer of that experience to the following strategic event. This transfer effect can be either positive or negative (Cormier & Hagman, 1987 in K. M. Ellis et al., 2011). Positive transfer occurs if there is similarity between the transfer source and the transfer target (Salomon & Perkins, 1989; Woodworth & Thorndike, 1901). When the acquirer recognizes common factors they can draw on prior experience and exploit this appropriately to the focal condition. In turn, generalizing similar experience to subsequent related acquisition continually improves performance. Conversely, when the focal acquisition is to novel or dissimilar from existing acquisition routines the transfer effect may be negative (Haleblian & Finkelstein, 1999; Hayward, 2002; Salomon & Perkins, 1989). However, both results are contingent when the conditions of appropriate generalization are met. The behavioural challenge for strategic decision makers is therefore being able to appropriately discriminate whether the deal is related or not, given the vast experience heterogeneity associated to strategic settings (i.e. industry, size). Furthermore negative transfer effects are more likely when managers incorrectly value acquisition routines and incorrectly use prior developed acquisition routines (Finkelstein & Haleblian, 2002; Haleblian & Finkelstein, 1999; Salomon & Perkins, 1989). Taken together, the impact of transfer effect depend upon the similarities between the focal acquisition and prior acquisition experience but is also contingent on the organizational behaviour to discriminate unrelated factors from organizational routines and act upon appropriately (Woodworth & Thorndike, 1901).

Literature gap and research question

Ample studies tested, based on the transfer learning theory premise, the impact of relatedness (similarity) between the focal acquisition and past acquisition experience (Finkelstein & Haleblian, 2002; Haleblian & Finkelstein, 1999; Hayward, 2002; Uhlenbruck et al., 2006). Relatedness is tested across several contingencies such as industry type, country, deal size specificity and technological proximity (Barkema et al., 1996; K. M. Ellis et al., 2011; Finkelstein & Haleblian, 2002; Haleblian & Finkelstein, 1999; Hayward, 2002; Jo et al., 2016; Nadolska & Barkema, 2007). Consequently, similarity implies that if the antecedent conditions is similar to past experience then the organizational behaviour should be to draw on prior accumulated acquisition experience generalize to the transfer target (Haleblian & Finkelstein, 1999). Empirical results indicate that firms can expect enhanced firm performance through value creation from related deals. However, different type of relatedness account for different effects on deal performance (Ahuja & Katila, 2001; Bower, 2001; Hayward, 2002). The current body of knowledge acknowledges the importance of relatedness and experience heterogeneity given the numerous studies on this field (Ahuja & Katila, 2001; Bower, 2001; K. M. Ellis et al., 2011; Finkelstein & Haleblian, 2002; Haleblian & Finkelstein, 1999; Hayward, 2002; Jo et al., 2016; Nadolska & Barkema, 2007, 2014). A growing stream of scholars argues that acquisition experience heterogeneity based on the strategic rational could affect post-merger performance and that understanding the (strategic) motives underlying them is important to successfully implement different types of acquisitions (Ahuja & Katila, 2001; Berkovitch & Narayanan, 1993; Bower, 2001; Hayward, 2002; Jo et al., 2016). Even though existing studies recognize this type of heterogeneity and address the importance, they did not empirically test the effects of strategic motives on the impact

between acquisition experience and deal performance. Based on insights from transfer learning theory I argue that the strategic motive may influence the appropriateness of exploiting or disregarding prior acquisition experience to the new acquisition and therefore could influence the positive effects of transfer. For example, surface similarities between accumulated acquisition experience of tangible assets (i.e. property plant and equipment) and the focal intangible assets (i.e. patents) may be evident. However, appropriate discriminating of their underlying acquisition routines or structural features for intangible assets such as a brand, technological knowledge and intellectual property is more challenging (Gerbaud & York, 2007; Hall, 1992, 1993; Mizik & Jacobson, 2003). Despite the valuable contributions of earlier research that tested the effects of deal experience heterogeneity in several contingencies (i.e. industry, country, size, technological proximity), I thus propose that we can extend our current insights if we also incorporate deal experience heterogeneity based on the strategic motive. Therefore, the aim of this study is to get a thorough understanding of the mechanisms of transfer theory in strategic settings, while incorporating the moderating effect of the strategic rationale on the impact between acquisition experience and deal performance. Derived from this, the research question of this study aims to answer the following:

What is the moderating effect of strategic motives, on the impact between acquisition experience and deal performance?

Non-linear relationship of acquisition experience and acquisition performance

Research on relatedness between acquisition experience and deal performance suggest that firms create value from their antecedent (focal) acquisition if they exploit

accumulated acquisition experience to similar subsequent situations (J. Ellis et al., 2011; Finkelstein & Haleblian, 2002; Haleblian & Finkelstein, 1999; Hayward, 2002; Jo et al., 2016; Nadolska & Barkema, 2007). Likewise, firms learn faster from their experience and become more skilled once they acquirer in similar conditions. For example, novice firms are only able to discriminate similarities in surface features, whereas experts are able to differentiate the surface and underlying structural features of the focal acquisition from previous experience (Haleblian & Finkelstein, 1999; Hayward, 2002; Nadolska & Barkema, 2007). Through the iterative process of organizational learning, firms gain contingent specific experience, learn from the consequences – reward or punishment – and subsequently accumulate the experience in acquisition routines (Haleblian & Finkelstein, 1999; Hayward, 2002; Levitt & March, 1988).

Initially, when a firm conducts their first acquisition they create a baseline with that acquisition specific experience (Haleblian & Finkelstein, 1999; Hayward, 2002). As a firm gradually develops from ‘novice’ to ‘expert’, managers at the strategic apex develop capabilities, that allow them to discriminate the underlying dissimilarities with less effort and in turn diminishes inappropriate generalization. However, several studies point out that too much homogenous acquisition experience is detrimental for deal performance and may lead the acquirer become prone to a competency trap (Hayward, 2002; Levitt & March, 1988). This competency trap takes place when firms achieve successful post-merger performance from inferior adopted routines and therefore accumulate this experience (Levitt & March, 1988). Because firms are unaware of this fact, the risk increases of inappropriate generalization and discrimination, which results in adverse transfer effects as well as negative deal performance (Haleblian & Finkelstein, 1999). Thus, performance is first negative due

to the fact that novice firms inappropriate generalize, followed by an upward in deal performance due to enhanced expertise and then suffer again a decrease in performance as they become prone to the competency trap (Haleblian & Finkelstein, 1999; Levitt & March, 1988). This relationship between similar acquisition experience and deal performance is therefore found to be non-linear u-shaped by several studies (Finkelstein & Haleblian, 2002; Haleblian & Finkelstein, 1999; Hayward, 2002; Jo et al., 2016; Nadolska & Barkema, 2007; Zollo & Singh, 2004). Therefore, scholars suggests that firms should not only exploit their related accumulated acquisition experience but also balance through exploring new dissimilar antecedent targets, which may prevent the firm to become cognitive inert (Hayward, 2002; Leonard-Barton, 1992). However, firms acquire capabilities and new resources specific to their industry for several strategic reasons such as market entering acquisitions to extend presence into new products and markets (Amburgey & Miner, 1992; Barkema et al., 1996; Bower, 2001; J. Ellis et al., 2011; Nadolska & Barkema, 2007; Reus, Lamont, & Ellis, 2016). Another rationale emphasizes strengthening a firm’s position in their market through exploiting industries, consolidating overcapacity and by rolling-up competitors in fragmented industries (Amburgey & Miner, 1992; Bower, 2001; Hayward, 2002; Hitt et al., 1997; Montgomery, 1994; Porter, 1987). Market-strengthening acquisitions are commonly executed and path dependent as they exploit the same industry that the firm operates, resulting in synergies from economies of scale and scope, creation and product development capabilities or technological knowledge (Ahuja & Katila, 2001; Bower, 2001; Hall, 1992, 1993). Consequently, exploiting similar acquisition experience to accessible antecedent markets enhances learning and deal performance up to a certain optimum but also hinders organizations to enrich their acquisition knowledge base. Still,

acquiring too many similar acquisitions is detrimental once the optimum is achieved. This makes the acquiring decision-makers prone to the competency trap and offsets their discriminative capabilities, which leads to inappropriate generalization (Hayward, 2002; Levitt & March, 1988). Consistent with existing studies, I therefore hypothesize that acquirers exploiting similar acquisition experience in the same four-digit SIC industry can expect a similar non-linear learning curve. This is formulated in the first hypothesis as:

H1: There will be an inverted u-shape relationship between the similarity of prior

acquisition experience and deal performance

Firm age

In this section I argue that the ability to discriminate by the acquiring firm, despite their acquisition experience, is affected by the age of the firm. Organizational learning theory stresses the importance of conducting similar subsequent acquisitions for novice firms (Haleblian & Finkelstein, 1999, 1999; Hayward, 2002; Nadolska & Barkema, 2007; Salomon & Perkins, 1989). As such, younger firms benefit from this experience homogeneity up to a certain point as it allows for gradually learning to discriminate surface and structural differences between the antecedent and focal acquisition (Hayward, 2002). Likewise, they develop associated core capabilities (i.e. negotiation, due diligence, integration) through a continuous process of translating, codifying and embedding experience into organizational routines (Levitt & March, 1988). Ample research on transfer theory effects in strategic settings confirm a positive relation for similar acquisition experience and deal performance; suggesting that once firms gradually evolve over time they develop their organizational

discriminative behaviour and become more skilled and experts in their industry (Al-Laham, Schweizer, & Amburgey, 2010; Dierickx & Cool, 1989; K. M. Ellis et al., 2011; Finkelstein & Haleblian, 2002; Haleblian & Finkelstein, 1999; Hayward, 2002; Jo et al., 2016; Woodworth & Thorndike, 1901). However, exploring and incorporating heterogeneous acquisition experience is a prerequisite to renew the knowledge base and acquisition capabilities, which is core to the organization. Firms that acquire too much of the same offset the benefits of core capabilities by becoming more cognitive inert, which in turn makes them core rigid (Leonard-Barton, 1992). This is especially important for organizations that operate in dynamic high-technological R&D intensive industries such as the biohigh-technological sector, pharmaceuticals or semi-conductor (Ahuja & Katila, 2001; Bena & Li, 2014; Datta & Roumani, 2015; Jo et al., 2016; Makri, Hitt, & Lane, 2010; Sears & Hoetker, 2014). Al-Laham, Schweizer & Amburgey (2010) elucidate in their study with support from empirical evidence, that there is a positive correlation between aging and acquisition experience, indicating that once companies become older they are more prone to administrative and core rigidities (Leonard-Barton, 1992). This reemphasizes the results from Haleblian & Finkelstein’s (1999) that novice firms are prone to inappropriate generalization, whereas older firms are susceptible to a competency trap (Levitt & March, 1988). In both situations transfer theory predicts negative implications for deal performance. It follows that the organizational behaviour to discriminate is not only determined by the number of prior acquisition experience but could also be affected by the firm age. To put differently, the underlying mechanism of transfer effects – the ability to discriminate – could be negatively affected by the age of the acquiring firm. I therefore expect that there is a two-way interaction between firm age and similarity of prior acquisition experience, as such that the

positive relationship between similar acquisition experience and deal performance is contingent upon the acquirer’s firm age. This is restated in the second hypothesis as:

H2: Firm age negatively moderates the impact of similar acquisition experience on

deal performance, as such that the positive effect of similar acquisition experience becomes weaker

Strategic motives

For many years scholars in the field of strategic management are interested in explaining differences in performance among firms. They agree that firms differ in terms of strategic focus by which they create a competitive advantage (Barney, 1986, 1991; Peteraf, 1993; Porter, 1987; Rumelt, 2003). Firms that operate in the same (four-digit) SIC industry can thus differ from competition in terms of strategic focus. They do so by exploiting internal resources and developing new capabilities (Leonard-Barton, 1992). However, resources and competency development is path-dependent and requires significant time to build over time (Dierickx & Cool, 1989). According to efficiency theory M&A is a means for firms to acquire these strategic renewal resources and (technological) capabilities in a fast way that allows the acquirer to gain synergies by using, internalizing or creating resources and capabilities (Haspeslagh & Jemison, 1991; Larsson & Finkelstein, 1999; Mizik & Jacobson, 2003). For example, a firm can decide to invest in a certain technology or alternatively acquire the already developed intangible asset from a competitor in the market. The strategic motive described in the firm’s official announcement statement signals which target resources are being sought and why (Gerbaud & York, 2007). Scholars argue that not all M&A’s are alike due to heterogeneity in motive-specific

experience and that being able to discriminate the underlying structural features between them is an important determinant to successfully implement a range of acquisitions (Ahuja & Katila, 2001; Berkovitch & Narayanan, 1993; Bower, 2001; Hayward, 2002; Jo et al., 2016). This is especially challenging for novice firms who are inexperienced in discriminating the underlying structural features between the antecedent and past conditions, which in turn is predicted to negatively affect deal performance (Haleblian & Finkelstein, 1999).

From an efficiency theory perspective on merger motives, studies explain that resources are a key variable in understanding decision making for value creation in acquisitions (Ahuja & Katila, 2001; Capron, Dussauge, & Mitchell, 1998; Capron, Mitchell, & Swaminathan, 2001; Karim & Mitchell, 2000). When it comes to categorizing resources the literature shows numerous developed categorical schemes such as, but not limited to, tangible and intangible resources (Hall, 1992, 1993), resources and competencies (Amit & Schoemaker, 1993; Grant, 1991) and property and knowledge-based resources (Miller & Shamsie, 1996). While each scholar emphasizes on a particular trait of the resource typology under scrutiny, the bottom line is that most of them share common ground (Gerbaud & York, 2007). For the purpose of this study the classification from efficiency theory is used. This theory considers that firms acquire tangible resources, intangible resources that are people dependent or intangible resources and capabilities that are not people dependent (Hall, 1992, 1993). Tangible resources are physical assets such as a manufacturing facility, production line or other types of equipment (Bower, 2001; Capron et al., 1998; Grant, 1991; Newmeyer et al., 2016). Research points out that firms that acquire products are signalling the importance of primarily acquiring tangible resources and capabilities

(i.e. manufacturing facility, plant and equipment)(Connelly, Certo, Ireland, & Reutzel, 2011; Newmeyer et al., 2016).

Intangible people dependent resources encompasses technological know-how of employees, distributors, suppliers as well as the beliefs, knowledge and shared mental minds captured in the organizational culture (Coff, 1999; Gerbaud & York, 2007; Hall, 1992, 1993). These types of intangibles can be classified as skills. Likewise, intangibles that are people independent can be classified as assets (Hall, 1992, 1993). This group entails intellectual property such as trademarks, patents, trade secrets, brand names, reputation as well as a firm’s network. Intangible resources and capabilities provide value for the acquirer if the firm has associated capabilities developed and embedded in the organization (Mizik & Jacobson, 2003; Newmeyer et al., 2016). Due to their value, rare, inimitable, non-substitutability and immobile traits they provide the firm revenues, cost-synergies and premiums, which are directly accretive to the organizational performance (Barney, 1986, 1991; Peteraf, 1993). Firms that acquire targets for reasons related to tangible and intangible resources accumulate this specific acquisition experience into their organizational routines (Levitt & March, 1988; Szulanski, 1996). Following the logic of behavioural learning and transfer theory, decision-makers of acquiring firms (novice and expert) should be able to appropriate differentiate similarities between tangible resources and intangible resources or capabilities and therefore experience positive transfer. To put differently, given that the surface features between tangible and intangible resources may be obvious to discriminate, I thus expect appropriate generalization and no significant differences in deal performance for both novice and expert firms. The third hypothesis therefore proposes that:

H3: Acquisitions driven by strategic motives to buy tangible or intangible resources

moderate the impact of similar acquisition experience on deal performance, as such that the positive effect of similar acquisition experience reports no significant

differences

When a manager of the acquiring firm buys human dependent intangible-resources, which are tacit and socially complex, the required level of acquisition routines will be different from that of human non-dependent intangible resources (Bower, 2001). Acquirers of knowledge-based resources are more susceptible to problems such as valuing, transferring and integrating these resources, due to their causal ambiguity, complexity and tacit traits, which increases uncertainty (Coff, 1999; Gerbaud & York, 2007; Larsson & Finkelstein, 1999; Ranft & Lord, 2002; Zander & Kogut, 1995). Challenges arise such as dealing with cultural compatibility and to prevent acquired skilled employees walk away after the post-merger integration, whereas brand resources or intellectual property are immobile and simply stay within the firm (Barney, 1986, 1991; Peteraf, 1993). The premise especially holds true in high-tech R&D intensive industries that acquire for reasons such as well-developed technological knowledge bases, significant creation and product development capabilities, brand resources or marketing capabilities (Ahuja & Katila, 2001; Bena & Li, 2014; Bower, 2001; Datta & Roumani, 2015; Hitt et al., 1997; Newmeyer et al., 2016; Sears & Hoetker, 2014). Firms in these industries acquire intangible resources from unrelated antecedent targets to enhance novel innovations as opposed to targets with complementary knowledge stocks, in which the acquirer benefits incremental innovations (Bena & Li, 2014; Makri et al., 2010; Sears & Hoetker, 2014). These different types of intangible resources along with their tacit traits, translate into

experience heterogeneity that challenges the ability of firms to recognize common factors between the focal acquisition and past condition and appropriately draw upon developed acquisition routines (i.e. negotiation, due diligence, integration and implementation)(Hall, 1992, 1993; Newmeyer et al., 2016; Ranft & Lord, 2002; Zander & Kogut, 1995; Zollo & Singh, 2004).

Studies from finance literature confirm these challenges among the type of intangible resources, reflected in stock market reactions. Firms release specific information (signals) about the strategic motive of the focal acquisition during the announcement statement to enhance information symmetry. Signalling theory explains that investors react to the information provided in the deal based on the type of resources under sale and analyse the potential value added (i.e. synergistic benefits)(Connelly et al., 2011; Gerbaud & York, 2007). To put differently, the stock market reactions reflect whether investors believe in the appropriateness of exploiting or disregarding prior acquisition experience to the new acquisition by the firm and if it result in positive transfer. Intangible people independent resources such as a brand or intellectual-property provide value for the acquirer if the firm has associated capabilities developed and embedded in the organization (Mizik & Jacobson, 2003; Newmeyer et al., 2016). Buying these capabilities is important for the acquirer given that the development is path-dependent (Dierickx & Cool, 1989). Due to their value, rare, inimitable, non-substitutability and immobile traits they provide the firm revenues, cost-synergies and premiums, which are directly accretive to the organizational performance (Barney, 1986, 1991; Peteraf, 1993). Insights from studies in the field of marketing and finance indicate that stock market reactions are positively associated for acquisitions driven by motives to buy intangible resources such as a strong brand because they are directly accretive to the acquirers firm performance (Newmeyer et al., 2016; Rego,

Billett, & Morgan, 2009). Likewise, the literature reports negative stock market reactions for acquisitions that involve knowledge-based resources such as skills, which are human dependent (Gerbaud & York, 2007). These reactions reemphasize that tacitness and social complexity can slow integration and hinder the positive transfer effects of prior acquisition experience (Larsson & Finkelstein, 1999; Szulanski, 1996). Due to the associated uncertainty, with regards to issues that affect post-merger deal performance such as integration, retention or reduced labour productivity investors react less positive to acquisitions motived to buy human dependent knowledge resources (Gerbaud & York, 2007).

Taken together, I expect a significant difference in deal performance based on the type of intangible resource sought, such that human independent resources has a positive moderating effect and human dependent resources a negative moderating effect on the impact of similar acquisition experience and deal performance. This is rephrased in the last hypothesis as:

H4: Acquisitions driven by strategic motives to buy human-independent resources

and human dependent resources moderate the impact of similar acquisition experience on deal performance, as such that the positive effect of similar acquisition

experience becomes stronger for human-independent resources and weaker for human-dependent resources

METHODOLOGY

This chapter elaborates the methodology undertaken for this study. The first section explains the sampling strategy and used source data. Subsequent sections, two, three, four and five discuss how the explanatory, moderating, independent and control variables are constructed, measured and what they will measure. Lastly, section six explains the statistical model that I use and how I deal effectively with spurious data.

Sampling strategy

The sample data for this study encompasses completed deals, based on secondary data, with an asset value larger than $10 million, undertaken between fiscal year 1989 and 2016 by publicly listed manufacturing firms with a four-digit standard industry classification (SIC) code ranging from 2800 and 2899 in North America. I select manufacturing firms to test the moderating effect of strategic motives on the transfer effect between acquisition experience and acquisition performance. This allows comparison with the existing body of knowledge (K. M. Ellis et al., 2011; Finkelstein & Haleblian, 2002; Haleblian & Finkelstein, 1999; Hayward, 2002; Jo et al., 2016). Therefore, firms that operate in the financial sector (i.e. banks) with SIC code 6000-6999 are excluded. I limit my sample to large deals because these tend to have a significant impact on market valuation (K. M. Ellis et al., 2011; Finkelstein & Haleblian, 2002; Haleblian & Finkelstein, 1999). Completed deals refer to acquisitions were the acquirer not only did a bid but also paid a cash premium or share exchange to get ownership over the target firm (Bena & Li, 2014). Secondary data on acquisition experience is gathered through M&A deal announcement data from the Standards & Poor (S&P) Global Market Intelligence database Capital IQ. In addition data is derived from the source Wharton Centre for Research in Securities

Pricing (CRSP) – Compustat Merged database. The combined CRSP – Compustat provides annual financial information, industry data, geographical presence and SIC codes for both target and acquirer firm. Market return data for the dependent variable, acquisition performance is derived from CRSP. This yielded data for a total of 151 large domestic U.S. deals. To determine the strategic motive of the acquisition I follow a similar approach applied by Gerbaud & York (2007) and Newmeyer et al. (2016) and gathered info by going through the statements released during the announcement date, 10K reports and SEC filings. I focus with my sample for strategic motives on the high-tech R&D industry with two-digit SIC code 28 as it encompasses the industry with the most acquisitions undertaken in the sample period as well as the highest transaction values. This is illustrated in table 1 in the appendix*. The industry focus in this study also matches the call from several scholars for a more holistic approach on transfer effects of learning in high-tech industries (Ahuja & Katila, 2001; Bower, 2001; Jo et al., 2016). After excluding missing values I end up with a unique database that is not used before, encompassing a total of 149 domestic acquisitions with a total transaction value of $687 billion, performed by 63 firms.

Dependent variable

The dependent variable in this study is acquisition performance, which captures the wealth that acquirers create with the focal acquisition (Hayward, 2002). This study aims to measure whether differences in acquisition experience heterogeneity from strategic motives has firm consequences (“reward” or “punishment”) for ex-post deal performance. A positive causal relationship between acquisition experience and deal performance suggest positive organizational behaviour in such a way that firms learn from prior conditions and generalize experience appropriately (Finkelstein &

Haleblian, 2002; Haleblian & Finkelstein, 1999; Hayward, 2002). To test if acquirers indeed learn from prior conditions by selecting the right focal deal, I use a conventional event study methodology that measures deal performance with cumulative abnormal returns (CAR) based on stock market reactions. The abnormal return is the change in the security price on the official announcement date and signifies the importance of the strategic event. It is calculated as the difference between the actual normal return and the predicted normal return (Finkelstein & Haleblian, 2002; Haleblian & Finkelstein, 1999). Using event study analysis assumes that capital markets are efficient and therefore stock prices and securities prices on the acquisition announcement date reflect all available, relevant information (Haleblian & Finkelstein, 1999; Liu, 2014). To put differently, the stock price reflect the potential synergistic effects that investors expect to accrue over time after the deal for the acquiring firm, based on the information (signals) provided from the official announcement statement (Connelly et al., 2011). Studies from finance and strategy literature report that there is evidence for an positive correlation between ex-ante measures of abnormal returns and ex-post measures for deal performance (Harrison & Godfrey, 1997; Healy, Palepu, & Ruback, 1992; Kaplan & Weisbach, 1992; Sirower, 1997). This proves that stock market may reasonably forecast the performance of the focal acquisition and strengthens the justification for choosing the dependent variable. To calculate the estimated abnormal return of the security around the official announcement date, I adopted an estimation model from 200 trading days before the event up to 46 days before the event. The 46 days interval has been chosen to remove the effects of news about the strategic event that may potentially leak prior to the official announcement and consequently affect the security price of the acquiring firm (Asquith, Bruner, & Mullins, 1983). I use an event window period starting one trading

day before the deal announcement to one day after the deal. This complies with previous conducted research (Finkelstein & Haleblian, 2002; Gerbaud & York, 2007; Haleblian & Finkelstein, 1999; Hayward, 2002; Newmeyer et al., 2016) and is commonly used in the finance literature (i.e. Liu, 2014).

Independent variable

Acquisition experience: consistent with preceding studies I operationalize the independent variable acquisition experience as the sum of the total number of prior acquisitions with an asset value greater than $10 million, undertaken by a firm starting from the first period of the sample (Barkema & Vermeulen, 1998; K. M. Ellis et al., 2011; Finkelstein & Haleblian, 2002; Haleblian & Finkelstein, 1999; Hayward, 2002; Jo et al., 2016; Nadolska & Barkema, 2007, 2014).

Target-to-target similarity: the similarity of prior acquisitions is proxied as the similarity between transfer target (focal acquisition) and transfer source (prior experience) defined at the four-digit industry level (Finkelstein & Haleblian, 2002; Haleblian & Finkelstein, 1999; Hayward, 2002). The measure is a proportion of prior acquisitions that match the focal four-digit SIC code. An acquisition is assigned with the score 1 if all prior acquisitions are within the same four-digit SIC code. In a similar vein prior acquisitions are labelled with 0 zero if no two acquisitions share a similar four-digit SIC code. Lastly, if one out of the two deals has the same code then a value of 0,5 is assigned. I then multiply these scores by a factor of ten so that their squared terms are used to model and test the non-linear inverted u-shape relationship as stated in hypothesis 1 (Hayward, 2002).

Moderating variables

Firm age: to determine the firm age of the acquirer, I establish a continuous variable, which measures the difference in the number of days between the official announcement date of the focal acquisition and the founding year.

Tangible and intangible resources: to assign the strategic motive to the focal acquisition, based on the type of resources sought (tangible and intangible), I follow the same methodological approach as Gerbaud & York (2007) and Newmeyer et al. (2016). I scrutinize each of the acquisition statements released during the announcement date, 10K reports and SEC filings and listed words belonging to the type of strategic motive. Table 2 shows an example excerpt from acquisition announcements along with the classification criteria. The take-over is classified as motivated to buy strategic tangible resources and capabilities if the announcement involves the words product portfolio, product pipeline, tangible assets, manufacturing facility or plant and equipment related (Capron et al., 1998; Grant, 1991; Newmeyer et al., 2016). Conversely, if the announcement involves the words brand assets, intellectual property, technological knowledge and drug discovery technology the acquirer signals to the stock market that the deal is motived to buy intangible resources (Gerbaud & York, 2007; Hall, 1992, 1993; Newmeyer et al., 2016). The keywords from the acquisition announcement are then coded and given the score 0 if the motive is driven to buy tangible resources and assigned 1 if the motive is to buy intangible resources.

Human dependent and human independent resources: to assign the strategic motive to the focal acquisition, based on the type of intangible resources and capabilities sought (human dependent and human independent), I establish in a similar vein a binary variable. The approach also entails reading deal announcements, newspapers, 10K

reports, SEC filings and scanning specific indicators that motivate the deal (Gerbaud & York, 2007). Based on the works from Gerbaud & York (2007) and Hall (1992, 1993) I classify human dependent resources by searching for motives that refer to human skills such as knowledge, drug discovery technology or culture. For human independent resources I used the typology from Newmeyer et al. (2016) and searched for signals in the official statement release that involved the words brand assets, intellectual property, trade secrets or patents. Subsequently the focal acquisition is assigned 0 for human dependent resources and 1 for human independent resources.

Table 2: Excerpts from acquisition announcements

Acquisition Motive Date Description

Any acquisition announcement that involved the words ‘tangible assets’, ‘manufacturing facility or plant & equipment related’, ‘product portfolio’ or ‘product pipeline’

Tangible asset November

18th_{, 1999}

US – Baxter International Inc. announced today that it is acquiring North American Vaccine, Inc. We’ll be able to capitalize on North American Vaccine’s extensive pipeline by bringing Baxter’s core competencies […] and operational discipline.

Tangible asset November

6th_{, 2006}

US – Abbott Park, ILL, and Kos Pharmaceuticals, Inc. today announced a definitive agreement […] “Kos Pharmaceutical is an excellent fit for Abbott bot scientifically and commercially” […] this acquisition expands Abbott’s presence in the lipid management market and provide several on-market and late-stage pipeline products.

Tangible asset November

24th_{, 2008}

US – Johnson & Johnson and Omrix Biopharmaceuticals, Inc., a fully integrated biopharmaceutical company that develops and markets bio surgical and immunotherapy products […] will be acquired. The acquisition of Omrix would provide ETHICON Inc. a J&J company […] to strengthen its presence in active, biological-based haemostats and convergent products for various surgical applications. Any acquisition announcement that involved the words ‘brand assets’, ‘intellectual property, ‘patent’, ‘technological knowledge’ or ‘drug discovery technology’.

Intangible asset January 8th,

2005

US – Proctor & Gamble Co. said Friday it will buy the Gillette Co. brand in a stock swap worth about 57 billion, creating the world’s largest consumer products company with 21 brands.

Intangible asset January 26th_,

2012

US – Amgen Inc., the world largest biotechnology company, said on Thursday it would pay $1.16 billion to buy Micromet Inc. MITI.O, in a deal that would give it access to a novel cancer treatment technology.

Intangible asset January 12st_,

2014

US – Alnylam Pharmaceuticals Inc., a leading RNAi therapeutics company, announced today the acquisition of Merck’s wholly owned subsidiary Sima Therapeutics, Inc., comprising intellectual property […] and other delivery technologies.

Control variables

Target-to-acquirer similarity: capital markets react on average more positive to related acquisitions than to unrelated acquisitions due to the expected synergies that accrue to the acquirer post-merger (Goold & Luchs, 1993; Haleblian & Finkelstein, 1999; Porter, 1987). Empirical evidence support that unrelated acquisitions (conglomerate acquisitions) lead to value destruction, whereas the opposite is also true (Goold & Luchs, 1993; Montgomery, 1994; Rumelt, 1974). It is important to mention that these effects differ per industry. Unrelated acquisition in high-tech R&D industries for example is found to spur the innovation process and output of novel products (Bena & Li, 2014; Lodh & Battaggion, 2015; Makri et al., 2010; Sears & Hoetker, 2014). To control for the market reactions, with regards to acquirer-to-target similarity, I construct a binary variable that measures industry relatedness based on the four-digit SIC codes. Thus, when the industry is similar I assign 1 and 0 if otherwise (K. M. Ellis et al., 2011; Finkelstein & Haleblian, 2002; Haleblian & Finkelstein, 1999; Hayward, 2002).

Relative acquisition size: research confirms that the ratio of acquirer size in relation to the target, positively relates to cumulative abnormal returns (Bena & Li, 2014; K. M. Ellis et al., 2011; Finkelstein & Haleblian, 2002; Haleblian & Finkelstein, 1999; Hayward, 2002; Netter et al., 2011). Value creation from acquisition tends to benefit larger firms more than their smaller peers. That makes this research prone to size bias. To control for acquisition size I use target-acquirer ratio size, measured in terms of target assets to acquirer assets, one year prior to the focal deal (Bena & Li, 2014; K. M. Ellis et al., 2011; Finkelstein & Haleblian, 2002; Haleblian & Finkelstein, 1999). Total assets are computed as total assets plus total liabilities as stated by the accounting principles on the firm’s balance sheet.

Financing of the deal: abnormal returns for the acquiring firm tend to be dependent upon the financing of the deal. Capital markets react more positive to cash offerings, whereas stock offerings on average results in negative reactions from investors around the announcement date (J. Ellis et al., 2011; Netter et al., 2011). The reason is that all-cash transactions make for simple equity value calculation. Next it represents the cleanest form of currency and certainty of value for all shareholders. Yet, cash financing triggers taxable events as opposed to the stock exchange, which are tax-free. To compensate for the taxes paid, acquiring firms must pay a higher deal price (Haleblian & Finkelstein, 1999). Deals financed with stock-offering signal overvaluation of the acquirer assets. The explanation is that, due to asymmetric information, the acquiring firm can have inside information that is not reflected in the stock-price. Therefore, given the different expected abnormal returns on the announcement date, the type of financing indicates biases. To control for this bias I measure this variable on an ordinal scale, indicating 0 for share-for-share exchange, 2 for all-cash transaction and 1 for a combination.

Deal size: the size of the focal deal under consideration has implications for firm performance, as such that cumulative abnormal returns on the announcement date are related to the deal size. Scholars found empirical evidence that unrelated size-specific experience, conform transfer theory predictions, hurts deal performance (K. M. Ellis et al., 2011). Firms with experience in large related acquisition benefit from generalizing past experience. However, transferring accumulated experience from small acquisitions to large antecedent situations tend to be more difficult due to the complexity of the deal. This calls out for specific competencies in acquisition routines such as post-merger integration. Furthermore, taking into account that value-creation increases per merger wave and that megamergers appear more often with the intention

to strategically realign or achieve industry consolidation to exploit economies of scale and scope, I therefore control for deal size. A binary variable measures the focal deal size as such that 1 is assigned for a transaction with a deal value larger than $100 Mio and 0 for deals between the range of $10 to $100 Mio.

Timing of prior acquisitions: research shows that both long intervals and short intervals between subsequent acquisitions are detrimental to transfer learning effects (Hayward, 2002). Acquisitions are not only very demanding and complex projects but are also accompanied with significant levels of stress for the dealmakers involved in the purchasing process. Therefore, given that development of competencies requires significant time, too short intervals hinder managers to enhance their acquisition competencies by learning from the antecedent experience (Dierickx & Cool, 1989; Levitt & March, 1988). Organizations learn from their established acquisition routines and people involved in the deal. However, as firm ages over time, consequently the chance arises that employees leave the firm (Coff, 1999; Gerbaud & York, 2007; Hayward, 2002). The literature shows that if firms do not codify this experience, they are less likely to draw appropriate inferences from these past conditions (Hayward, 2002; Szulanski, 1996). To control for the timing between acquisitions I construct a continuous variable that measures the number of days between the focal and previous acquisition (Hayward, 2002).

Merger cyclicality: the timing of the deal can have substantial impact on the abnormal returns on the announcement date. In general, acquisitions appear in cycles of seven years, characterized by different aspects that affect abnormal return for acquiring firms. Each period is spurred by macro-economic conditions (i.e. interest rates, globalization, equity prices) that affect acquisition activity (Haleblian & Finkelstein, 1999; Montgomery, 1994). Examples of particular waves and their focus are

divestments, strategic megamergers, cross-border acquisitions and horizontal megamergers with the aim to realign or achieve global consolidation of manufacturing industries (Barkema & Vermeulen, 1998; Goold & Luchs, 1993; Hébert, Very, & Beamish, 2005; Nadolska & Barkema, 2007; Reus et al., 2016). If capital markets tend to react more positive to related (horizontal) acquisitions and divestments of unrelated business lines, reflected in higher abnormal returns for acquirers, then merger cyclicality is prone to a bias.

While research stresses out the potential impact of period effects on abnormal returns, controlling for merger waves is challenging from a methodological standpoint in determining a given year for a particular cycle. I therefore did not control for the effect of merger cyclicality. As an alternative I did test for period effects using a more common and easier approach in the literature with dummy variables for all the particular years in the sample. The period under scrutiny is assigned 1 and the rest 0. In line with previous studies the analysis did not yield significant results in each model and is therefore not reported (K. M. Ellis et al., 2011; Finkelstein & Haleblian, 2002; Haleblian & Finkelstein, 1999; Hayward, 2002).

Statistical model

Consistent with existing studies on transfer effects in strategic settings I run an Ordinary Least Squares (OLS) regression analysis (K. M. Ellis et al., 2011; Finkelstein & Haleblian, 2002; Haleblian & Finkelstein, 1999; Hayward, 2002). Attention should be paid for biases in the data using OLS, which can come from outliers, significant variance in the standard errors and interrelation (dependence) of observations. Analysis among the used variables in the models indicates the presence of outliers for the dependent variable deal performance. To diminish the effects of

these extreme values, I apply the strategy of winsorizing to the cumulative abnormal returns (CAR). Because the unique dataset that I established is smaller compared to similar studies on transfer learning theory, I argue with support from the literature that it is better not to remove valuable observations through trimming/truncation but rather modify the outliers accordingly, using the winsorizing methodology (K. M. Ellis et al., 2011; Finkelstein & Haleblian, 2002; Haleblian & Finkelstein, 1999; Hayward, 2002; Jo et al., 2016; Nadolska & Barkema, 2007). Following the literature on statistics I modified the CAR data subject to spurious outliers at the level 1% and 99% (Hampel, Ronchetti, Rousseeuw, & Stahel, 2005). This means that the extreme values

below the 1st percentile are set at the 1st percentile and likewise data above the 99th

percentile are set at the 99th _percentile.

The presence of autocorrelation and heteroskedasticity can invalidate the statistical significance of the effects as well as the inferences derived from the outcomes. Therefore I tested the linear regression models for homoskedasticity with the White test. The models 1, 2, 4 and 4 failed the nil hypotheses at the significant level of p < 0.05. To deal effectively with these concerns I applied a regression analysis with robust standard errors to these models (Hampel et al., 2005; Verardi & Croux, 2009). In a similar vein I tested each model using the Breusch-Godfrey serial correlation Lagrange multiplier test for autocorrelation (Breusch, 1978). This was found to be not presence in all of the executed statistical models and therefore no alternative test had to be applied. The models in this study are analysed with the STATA program.